Scatter/Gather Searching and Browsing with Bing API

Scatter/Gather Searching and Browsing

Live demo at:

The Scatter/Gather system integrates Bing search API for search and the Weka machine learning package for text clustering. After the user enters a search query, the system sends the query to Bing, retrieves the first 200 results, and clusters the results into a number of groups (7 clusters by default). The classic search system also integrates Bing search API but it directly presents search results to the user without clustering. 

With the Scatter/Gather system, users can select the clus- ters and the desired number of clusters in the next iteration. Documents in the selected (checked) clusters will be gath- ered and then clustered into the desired number of clusters. The user can continue searching/clustering iterations until they find what they need.

Figure is a screen shot of the Scatter/Gather interface. In the first iteration, seven clusters are identified and pre- sented to the user. A cluster description shows the number of member pages assigned to the cluster and top 10 key words generated from these results. An advanced option panel in the upper right corner enables the users to select the desired number of clusters that will be shown in the next iteration, ranging from 2 to 10.

Initially, the first two member pages of each cluster are shown. If there are more than two results in a cluster, the user can click “more” to get more results, and click “less” to show fewer items. When a potentially relevant page is presented, the user may click the link to view the page in a popup window.